This transform op runs a pass on the target op.
Details
Details
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
Comment Actions
I would consider evolving this into constructors of the entire pass pipeline rather than eager runners. So it can benefit from, e.g., parallelization and analysis preservation.
mlir/lib/Dialect/Transform/IR/TransformOps.cpp | ||
---|---|---|
384–398 | Can this be factored out? Something similar exists for "apply patterns". |
Can this be factored out? Something similar exists for "apply patterns".